mysql_skill

Overview

This skill executes safe, read-only SQL queries against one or more configured MySQL databases. It helps explore schemas, list tables, and run SELECT-style queries while preventing any write or destructive operations. Connections are chosen intelligently from descriptive entries so the right database is used for a given intent.

How this skill works

The skill reads a connections.json file containing named database entries with host, credentials, and optional SSL settings. It enforces a read-only session, validates that only allowed statements (SELECT, SHOW, DESCRIBE, EXPLAIN, WITH) are executed, and rejects multi-statement or write queries. Result sets are capped (row count, column width) and queries use a timeout to protect memory and execution resources.

When to use it

• Run ad-hoc SELECT queries against production or staging databases
• Explore database structure with table lists and schema dumps
• Validate data, inspect recent records, or sample rows for analysis
• Check specific tables or fields before making application changes
• Quickly confirm counts, joins, or aggregated results without risking writes

Best practices

• Store credentials in connections.json with file permissions set to 600
• Provide clear, concise descriptions for each connection to enable auto-selection
• Use --list, --tables, or --schema to inspect structure before composing queries
• Always include LIMIT for wide scans; the skill enforces a maximum row cap
• Avoid long-running full-table scans; rely on indexed predicates where possible

Example use cases

• List configured databases and pick the one containing users or orders
• Show schema for the orders table to verify available columns before querying
• Run SELECT queries to sample the latest 100 rows from a transactions table
• Describe a table to get column types and nullability for data-mapping tasks
• Run EXPLAIN on a query to review index usage and performance characteristics